It matters how you ask the question.
The Pew Study mentioned below tested whether it makes a difference if you ask people a recall question (Who is the president of Russia) versus can you tell me which of these guys is the president of Russia (answer: Putin). Scroll down to the sixth table. The results are simple: 36% got the recall question right. They could pull out of their heads (or some other body part) Putin's name. Give them a multiple guess and 60% nail the recognition question.
Hey, we're not so stupid after all.
In my own research there is support for the idea that television is better suited for recognition-based knowledge while print is better suited for recall-based knowledge. I'll explore this more at some other time, but it says a great deal about how we measure knowledge and whether our tests mean what we think they mean.